M. Parimala

Recent trends have shown a widespread increase in the landscape of digital healthcare (i.e., Healthcare 4.0) services, such as personalized healthcare, intelligent rehabilitation, telemedicine, and smart diet management, among others. These healthcare services are based on a variety of diverse requirements. Fulfilling these requirements require proactive intelligent analytics and self-sustainability of networks. Self-sustainability enables the operation of a network with minimum possible interaction from the end-users/network operators, whereas proactive intelligent analytics enables efficient management of resources in response to users’ requests. To enable healthcare 4.0 with proactive online analytics and self-sustainability, one can leverage digital twins. In this article, we present an overview and recent advances of digital twins for healthcare 4.0. An architecture of digital twins for healthcare is also proposed. Furthermore, we present several use cases of digital twins. Finally, we present open research challenges with possible solutions.

Federated learning (FL) is a recent development in artificial intelligence, which is typically based on the concept of decentralized data. As cyberattacks are frequently happening in the various applications deployed in real time, most industrialists are hesitating to move forward in adopting the technology of the Internet of Everything. This article aims to provide an extensive study on how FL could be utilized for providing better cybersecurity and prevent various cyberattacks in real time. We present an extensive survey of the various FL models currently developed by researchers for providing authentication, privacy, trust management, and attack detection. We also discuss few real-time use cases that have been deployed recently and how FL is adopted in them for preserving privacy of data and improving the performance of the system. Based on the study, we conclude this article with some prominent challenges and future directions on which the researchers can focus for adopting FL in real-time scenarios.

Human–computer interaction (HCI) and related technologies focus on the implementation of interactive computational systems. The studies in HCI emphasize on system use, creation of new techniques that support user activities, access to information, and ensures seamless communication. The use of artificial intelligence and deep learning-based models has been extensive across various domains yielding state-of-the-art results. In the present study, a crow search-based convolution neural networks model has been implemented in gesture recognition pertaining to the HCI domain. The hand gesture dataset used in the study is a publicly available one, downloaded from Kaggle. In this work, a one-hot encoding technique is used to convert the categorical data values to binary form. This is followed by the implementation of a crow search algorithm (CSA) for selecting optimal hyper-parameters for training of dataset using the convolution neural networks. The irrelevant parameters are eliminated from consideration, which contributes towards enhancement of accuracy in classifying the hand gestures. The model generates 100 percent training and testing accuracy that justifies the superiority of the model against traditional state-of-the-art models.

The rapid growth of data in Spatio-temporal datasets collected from several domains such as climate science, Health Science, Sensor networks and telecommunication systems has created a need for Spatio-Temporal clustering methods to extract and analys